Cohere has released North Mini Code, a 30-billion-parameter Mixture-of-Experts model with 3 billion active parameters, made available on Hugging Face under the Apache 2.0 license. It is the first in a new family of Cohere models and is purpose-built for agentic software engineering tasks, scoring 33.4 on Artificial Analysis' Coding Index and reportedly outperforming several larger open-source models in its class. The model was developed using a multi-scaffold training approach and a post-training pipeline combining supervised fine-tuning with reinforcement learning to support coding agents such as OpenCode.

Cohere has launched North Mini Code, its first developer-focused model, a 30B-parameter Mixture-of-Experts model with 3B active parameters designed for agentic software engineering tasks. Optimized for complex workflows and high-quality code generation, it outperforms competing models in its size class and is available on Hugging Face under the Apache 2.0 license. The model was trained using multiple scaffolds and features a two-stage fine-tuning process followed by reinforcement learning for coding excellence.